Associations between fungal species and water damaged building materials

AEM Accepts, published online ahead of print on 29 April 2011 Appl. Environ. Microbiol. doi:10.1128/aem.02513-10 Copyright 2011, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved. 1 Associations between fungal species and water damaged building materials 2 3 4 Birgitte Andersen 1 *, Jens C. Frisvad 1, Ib Søndergaard 1, Ib S. Rasmussen 2 and Lisbeth S. Larsen 2 5 6 7 8 1 Center for Microbial Biotechnology, DTU Systems Biology, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark. 2 Danish Technological Institute, Gregersensvej 1, DK-2630 Taastrup, Denmark 9 10 11 12 13 14 15 16 * Corresponding author Mailing address: CMB, DTU Systems Biology, Søltofts Plads, Building 221, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark. Phone: (+45) 4525 2726 Fax: (+45) E-mail: ba@bio.dtu.dk Running title: Fungi on water-damaged building materials 17 18 Key words: identification, health, moisture, mycotoxins, sampling methods 1

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Abstract: Fungal growth in damp or water damaged buildings worldwide is an increasing problem, which has adverse effects on both the occupants and the buildings. Air sampling alone in moldy buildings does not reveal the full diversity of fungal species growing on building materials. One aim of this study was to estimate the qualitative and quantitative diversity of fungi growing on damp or water damaged building materials. Another was to determine if associations exist between the most commonly found fungal species and different types of materials. More than 5,300 surface samples were taken by means of V8 contact plates from materials with visible fungal growth. Fungal identifications and information on building material components were analyzed using multivariate statistic methods to determine associations between fungi and material components. The results confirmed that Penicillium chrysogenum and Aspergillus versicolor are the most common fungal species in water damaged buildings. The results also showed Chaetomium spp., Acremonium spp., and Ulocladium spp. to be very common on damp building materials. Analyses show that associated mycobiotas exist on different building materials. Associations were found between 1) Acremonium spp., P. chrysogenum, Stachybotrys spp., Ulocladium spp., gypsum and wallpaper, 2) Arthrinium phaeospermum, Aureobasidium pullulans, Cladosporium herbarum, Trichoderma spp., yeasts, different types of wood and plywood and 3) A. fumigatus, A. melleus, A. niger, A. ochraceus, Chaetomium spp., Mucor racemosus, M. spinosus, concrete and other floor related materials. Results can be used to develop new and resistant building materials, relevant allergen extracts and to help focus research on relevant mycotoxins, MVOCs and micro particles released into the indoor environment. 2

40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 Introduction Most water damage indoors is due to natural disaster (e.g. flooding) or human error (e.g. disrepair). Water can seep into a building as a result of melting snow, heavy rain, or sewer system overflow. Water vapor can be produced by human activities like cooking, laundering, or showering and then condense on cold surfaces like outer walls, windows, or furniture. Damp or water damaged building materials are at high risk of fungal growth (mold growth), possibly resulting in health problems for the occupants and the deterioration of the buildings. The water activity ((a W ) a W 100 = % relative humidity at equilibrium) of a building material is the determining factor for fungal growth and varies with the temperature and the type of material (27). The longer a material is over 0.75 in a W the greater the risk of fungal growth (49), though different fungi have different a W preferences (11). Some filamentous fungi can grow on a material when a W is as low as 0.78 (26), while others can survive three weeks at 0.45 a W (30). The severity of indoor dampness varies with the climate, but WHO (51) estimates that in Australia, Europe, India, Japan, and North America, dampness is a problem in 10-50 % of the buildings, and Sivasubramani et al. (42) estimate that fungal growth is a problem in 15-40 % of North American and Northern European homes. The negative health effects of damp building materials and fungal growth in homes, institutions and workplaces have been reported in many publications including the WHO guidelines, Dampness and Mould (51), which concluded that there is sufficient epidemiological evidence to show that occupants of damp or moldy buildings are at increased risk of respiratory problems, respiratory infections and the exacerbation of asthma. The causality between fungal exposure and development of type I allergy has been proven (18), but clinical evidence linking specific fungal spores, hyphal fragments, and/or metabolites to particular health complaints is lacking. The symptoms reported by occupants in moldy buildings are many and diverse (20, 23) as are the fungal species found on moldy building materials (14, 19). The fact that some people are 3

64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 hypersensitive to fungi while others do not react at all further complicates the issue. Detection and species identification of all fungi present in a moldy building is the first step towards resolving the cause and effect of building related illness (sick building syndrome) so the choice of sampling methods is essential. Air and dust samples have been taken in order to associate fungal exposure and health problems (e.g. 10, 17, 48), but no conclusive links have been found. This may be because spore liberation from a surface is sporadic (42) and spore distribution in the air is random (21). Toxic fungi (e.g. Stachybotrys spp. and Chaetomium spp.) growing on damp building materials, do not readily become airborne and/or lose their culturability soon after liberation (21, 29, 35, 46) and may therefore not be detected during air or dust sampling. Correct species identification of the fungi is also important since new research has indicated that species specific metabolites, like atranone C produced by Stachybotrys chlorohalonata (38), are cytotoxic, or immunotoxic or induce inflammatory responses when inhaled (24, 33, 34). The purpose of this study was therefore to estimate the qualitative and quantitative diversity of fungi growing on damp or water damaged building materials. The study was based on more than 5,300 surface samples taken by means of V8 contact plates from building materials with visible fungal growth in Denmark and Greenland. The aim was to determine if there exists an association between the most common fungi found and particular types of water damaged building materials. 81 82 83 84 85 86 87 Materials and Methods Sample collection and treatment Samples from building materials with visible fungal growth were taken by means of 65 mm contact plates (VWR international) containing V8 agar with antibiotics (200 ml Campbell's original V8 100% Vegetable Juice (www.herbsgardenshealth.com or www.amazon.com), 3 g CaCO 3 (Merck), 20 g agar (VWR international) and 800 ml water. Penicillin (100,000 IU/l (Sigma)) and 4

88 89 90 91 92 93 94 95 96 97 98 99 streptomycin (1 g/l (Sigma)) were added after autoclaving (12)). The plates were subjected to fungal analysis at the Mycological Laboratory (ML) at The Danish Technological Institute (DTI). Samples were collected from June 2005 to February 2008 and originated from private residences (houses, apartments and holiday cottages), private businesses (shops and offices) as well as from public buildings (kindergartens, schools and offices) from all parts of Denmark and Greenland. Samples were taken from buildings where either professional building inspectors had reported visible fungal growth or water damage or occupants had contacted DTI with self-reported fungal or health problems. Several samples may have been taken from the same building, but only one sample was taken from each damage site. Approximately 75 % of all samples were taken on site by the building inceptors by means of contact plates and mailed overnight to ML. The remaining 25 % were moldy building materials sent to ML by the occupant after thorough instruction. The materials were then sampled by ML by means of contact plates. 100 101 102 103 104 105 106 107 108 109 110 111 Fungal identification Identification of fungi in a sample was done directly on the V8 contact plates after 7 days of incubation at 26 C in darkness. Whenever possible, fungi were identified to species using direct microscopy and identified according to Domsch et al. (7), de Hoog et al. (6), and Samson et al. (37). Fungi present in the sample were determined qualitatively (taxon present) and quantitatively (number of colonies). Since different Penicillium species can be difficult to identify on V8 medium, special attention was given to this genus in the spring of 2010. DTI randomly selected 80 V8 contact plates with Penicillium growth and delivered them to Center for Microbial Biotechnology (CMB) at the Technical University of Denmark (DTU). At CMB all different Penicillium colonies on each plate were isolated resulting in 120 Penicillium cultures. After transfer to CYA for purity control, the 5

112 113 isolates were inoculated onto CYA, MEA, YES and CREA and identified to species after 7 days at 25 C in the dark by methods reported by Samson et al. (38). 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 Data compilation The samples were evaluated on the basis of the reliability of information on material type and fungal identification. Each sample contained information on type of building (e.g. private home), type of water damaged construction (e.g. wallpaper on plaster, outer wall), qualitative fungal analysis (e.g. "A. niger (dominant), Chaetomium sp.") and quantitative fungal analysis (e.g. "30 A. versicolor, 1 Ulocladium sp."). The information of "type of water damaged material" was divided into categories. If an entry contained two or more components, it was split into component categories (e.g. "painted wallpaper on plaster" into "paint", "wallpaper" and "plaster"). The sample set containing qualitative data was transformed into a binary matrix consisting of 5,532 samples in rows and 51 different component categories and 57 fungi in columns. Fungi and component categories that constituted less than 0.5 % of the total 5,532 samples were deleted in order to minimize analytical noise (e.g. Karlit ceiling tiles, Botrytis cinerea, Doratomyces spp. or Epicoccum nigrum). Samples with growth of dry or wet rot fungi (Serpula lacrymans or Coniophora puteana, respectively) were also deleted. This resulted in a qualitative (binary) matrix (Matrix A) with 5,353 valid samples where the association between 30 building material components and 42 fungi was unambiguous. A similar process was repeated on the original sample set to extract the samples with quantitative data resulting in a matrix (Matrix B) with 4,241samples, 25 building material components, and 41fungi. 133 134 135 Multivariate statistics Matrices A and B were analyzed by Principal Component Analysis (PCA) using the program 6

136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 Unscrambler v9.2 (CAMO Process A/S, Oslo, Norway). PCA is a bilinear modeling method giving an interpretable overview of the main information. All variables (components and fungi) were standardized (X Average/Sdev), thus giving all the variables the same chance to influence the estimation of the components. In PCA proximities among the objects were judged using Euclidean distances and among the variables using covariance (or correlation) since the variables have been standardized. The information carried by the original variables was projected onto a smaller number of underlying ( latent ) variables called principal components. The data in Matrices A and B were then converted into two contingency tables of observed occurrences, where either the fungal count or the number of colonies for each fungus was summarized on each material. This resulted in two tables, contingency Table A based on the qualitative data (5,353 samples summarized into Table A [42 rows with fungi and 30 columns with materials]), and contingency Table B on the quantitative data (4,241 samples summarized in Table B [41 rows with fungi and 25 columns with materials]). From the contingency tables of observed occurrences, predicted values were calculated for a particular fungus on a particular material: (sum of counts/colonies for fungus A on all materials sum of counts/colonies of all fungi on material B)/(sum of counts/colonies of all fungi on all materials). For example, 7,452 Ulocladium colonies were counted in total, 19,100 fungal colonies was counted on all wallpaper samples and 366,304 fungal colonies were counted in total, giving a predicted number of Ulocladium colonies on wallpaper of (7,452 19,100)/366,304 = 388 compared with the observed number of 1,208 Ulocladium colonies counted on all wallpaper. Contingency Tables A and B were then analyzed by Correspondence Analysis (CA) using the program NTSYS version 2.21c (Exeter Software, Setauket, NY, USA) (15). Chi-square distances were used to judge proximities for both the row and for the column variables. The data in Matrix A was also converted into a fungal species distance matrix where the 7

160 161 162 163 164 165 count for each of the 42 fungi was summarized on the other 41 fungi. This was done to analyze if any of the fungal species co-occurred independently of material preferences. This resulted in Matrix C, a qualitative 42 42 symmetric matrix, which was then analyzed by Principal Coordinate (PCO) analysis using NTSYS v. 2.21c. Matrix C was double-centered and an eigenvector analysis was performed. The Correlation coefficient was used and a minimum spanning tree analysis was superimposed upon the OTUs in the PCO score plot (43). 166 167 168 169 170 171 172 173 174 175 Results Building materials Table 1 shows the building material components that were most often affected by fungal growth. As can be seen, plaster and concrete were the material components most likely to support fungal growth of the total material components. Together with wood, wallpaper, and gypsum, they constitute ca. 80 % of materials and construction parts damaged by dampness, condensation, or liquid water. The other 18 building materials that occurred with less than 2% were Masonite, cardboard, gas concrete, glue, wood-wool cement, bitumen, paper, vapor barriers, carpets, cork, MDF fiber board, vinyl, felt, grout, filler, Eternit, textiles, and tar-treated materials. 176 177 178 179 180 181 182 183 Fungi The raw data showed that 45 fungal genera or species in total were identified on the samples of water damaged building materials. Table 2 shows the qualitative and the quantitative presence of fungi on 5,353 and 4,241 samples, respectively. As can be seen Penicillium was the most dominant fungal genus (3,720 counts and 114,143 colonies) on water damaged building materials. The most dominant fungal species was Aspergillus versicolor (1,421 counts and 44,665 colonies). Together with Chaetomium spp., Acremonium spp., Ulocladium spp. and Cladosporium sphaerospermum 8

184 185 186 187 188 189 190 191 192 193 they constituted the most frequently detected fungi on damp or water damaged building materials. The other 15 fungi that occurred with less than 1% were Phoma spp., Paecilomyces lilacinus, A. ustus, Arthrinium phaeospermum, A. melleus, Alternaria spp., Scopulariopsis brumptii, Verticillium albo-atrum, A. flavus, P. variotii, A. sydowii, Absidia spp., Gliocladium spp., Guehomyces pullulans (syn: Trichosporon pullulans) and Aureobasidium pullulans. The isolation and identification of 120 Penicillia from 80 water damaged building materials (not the same samples as the above) showed that between 70 and 75 % of all Penicillium isolates were identified as P. chrysogenum, while P. brevicompactum, P. corylophilum, P. crustosum, P. olsonii, P. palitans and P. solitum constituted the last 25-30 % and were found in almost equal amounts. 194 195 196 197 198 199 200 201 202 203 204 205 206 207 Associations between fungi and building materials The result of a Principal Component analysis (PCA) of the qualitative (binary) data (Matrix A: 5353 samples 72 variables [42 fungi and 30 material components]) is shown in Fig. 1. The result of the PCA of the quantitative data (Matrix B: 4241 samples 66 variables [41 fungi and 25 material components]) gave a very similar result and is not shown. The first four PCA axes described 3%, 2%, 2% and 2% of the variation in both Matrix A and Matrix B. By plotting the first two principal component axes (PC1 against PC2), the interrelationships between all variables (fungi and material components) can be seen. The plot in Fig. 1 shows the qualitative associations between the different fungi, between the different components of building material and between fungi and material. The more often two fungal species occurred together in the same sample, the closer they are together in the plot: A. tenuissima, C. herbarum, Rhodotorula mucilaginosa and other yeasts together with A. pullulans, Fusarium spp. Trichoderma spp. and A. phaeospermum are often found together on different types of water damaged wood and therefore lie close together. The same is seen for the 9

208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 different building material components: Plaster, wallpaper, and painted surfaces co-occur in the plot in Fig. 1, because most Danish houses have brick walls leveled with plaster, coated with wallpaper and then painted. As can be seen, Acremonium spp., P. chrysogenum, Stachybotrys spp. and Ulocladium spp. often occur together and are highly associated with water damaged walls with painted wallpaper or glass fiber. On the other hand, Chaetomium spp., Penicillium spp. and different Aspergillus species were often found on water damaged concrete. Figure 1 also shows that A. versicolor, Calcarisporium arbuscula and Sporothrix spp. are placed opposite wood (negative correlated) in the plot meaning that A. versicolor, C. arbuscula and Sporothrix spp. occurred rarely on wood. The same negative correlation can be seen for wallpaper and A. niger, concrete and C. sphaerospermum, and plywood and A. ochraceus. Fungi or components lying close to the centroid in the plot occur infrequently (< 4%) and have very loose or little association with each other. Matrices A and B were converted to contingency Tables A and B based on the qualitative data (Table A: 42 rows with fungi and 30 columns with materials), and quantitative data (Table B: 41 rows with fungi and 25 columns with materials), respectively. Table 3 is an extract of contingency Tables A and B and shows the distribution of each of 17 fungi on 7 different materials. The table gives the qualitative and quantitative frequencies and distributions (in %) of fungi on different building materials, where the sum of all counts or colonies of one fungal species on all 30 or 25 materials constitute 100%. Some fungi are overrepresented on some materials (row values in bold) compared to the statistically predicted value, if the fungi were randomly or evenly distributed on all materials. For example, Sporothrix spp. have a strong association with plaster; i.e. 39% (162 counts) of all material samples with Sporothrix spp. were plaster and 37% (4,118 colonies) of all the Sporothrix spp. colonies counted were found on plaster samples. However, had Sporothrix spp. been randomly distributed, the statistically predictive values would have been 81 counts and 1,865 colonies, respectively. Sporothrix spp. also have an association with concrete where 18% of all 10

232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 concrete samples were colonized with Sporothrix spp. and 21% of all Sporothrix spp. colonies originated from concrete samples. However, Sporothrix spp. are rarely found on plywood (0 and 2%, respectively). Similar strong associations were found between Stachybotrys spp. and gypsum, wallpaper and glass fiber and between Ulocladium spp. and wallpaper, plaster and gypsum. Phoma spp. also had a preference for glass fiber. On water damaged wood A. tenuissima, C. herbarum, Trichoderma spp. and yeasts were the associated fungal species, while Mucor racemosus, A. ochraceus and A. niger were associated with wet concrete. It can also be seen that the most dominant fungal genus in water damaged buildings, Penicillium, was found to be evenly distributed on all the examined building materials, and without any pronounced preference. Likewise, some fungal species are underrepresented on some materials. For example, A. versicolor had a low association with plywood; i.e. 1% (20 counts) of all material samples with A. versicolor were plywood and 1% (776 colonies) of all the A. versicolor colonies counted were found on plywood samples. However, had A. versicolor been randomly distributed, the statistically predictive values would have been 41 counts and 4,307 colonies, respectively. Table 3 also shows that A. niger, A. ochraceus and M. racemosus were not associated with wallpaper and that C. sphaerospermum and yeasts were uncommon on concrete. Table 4 is also based on contingency tables A and B, but shows the distribution of 17 fungal species on each of the 7 different materials. The table gives the qualitative and quantitative occurrences and distributions (in %) of fungi on different building materials, where the column sum of all counts or colonies of all the 42 fungal species on one material constitute 100%. As can be seen Penicillium spp. is the dominant genus and can be found on all types of materials with almost the same frequency (27-30% quantitatively and 27-46% quantitatively) corresponding to the overall occurrence seen in Table 2. A. versicolor and Acremonium spp. are also very dominant and found evenly on most materials, except in the case of A. versicolor, where the distribution on plywood is 11

256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 underrepresented (5 and 6%) compared to its overall occurrence. The result of a Correspondence Analysis (CA) based on the quantitative data (contingency Table B: a 41 25 table based on Matrix B) is shown as a biplot in Fig. 2. The first four CA axes described 20.74%, 17.85%, 14.85% and 9.88% of the variation in the data for contingency Table B. The result of the CA of the qualitative data (contingency Table A: a 42 30 table based on Matrix A) gave a very similar result with the first four CA axes describing 21.54%, 17.06%, 12.02% and 8.37% of the variation in contingency Table A and is not shown. Figure 2 shows the quantitative associations between building material and fungi. As can be seen, wallpaper/glass fiber, gypsum, Stachybotrys spp. and Ulocladium spp. are strongly associated. Likewise, plywood is closely associated with A. phaeospermum, R. mucilaginosa and Trichoderma spp., while A. ochraceus, A. sydowii and M. racemosus are closely associated with each other and loosely associated with A. fumigatus and Mucor spp. and concrete, glue and cork. Being quite close to the centroid of the plot, Penicillium spp. is associated with most of the materials, while A. versicolor is more associated with gas concrete, cork, vinyl and glue than with wallpaper, grout/tile wood or plywood. 2-dimensional plots of both PCA and CA can sometimes make variables (i.e. fungi) appear more associated than they really are. A minimum spanning tree can be a good control whether particular variables (i.e. fungi) are indeed as close as their similar score or loading values suggest. To test the associations in Figs. 1 and 2, a Principal Coordinate (PCO) analysis with an overlain minimum spanning tree of Matrix C (Matrix C: 42 fungi 42 fungi) was made. The PCO analysis described 42.55%, 23.58%, 8.36% and 7.22% of the variation in the data on the first four axes, so these four axes described a larger proportion of the variance than expected using the broken stick model (10.30%, 7.92%, 6.73% and 5.94). The PCO analysis showed co-occurrence and strong association between 1) Acremonium spp., P. chrysogenum, Stachybotrys spp. and Ulocladium spp., 2) Trichoderma spp. A. tenuissima, C. herbarum, T. pullulans and Fusarium spp., 3) yeasts, 12

280 281 282 283 284 285 286 287 288 289 290 291 292 Gliocladium spp., A. phaeospermum and A. pullulans, 4) A. sydowii, A. ochraceus, M. racemosus, A. melleus, A. niger, M. spinosus, A. fumigatus, Chaetomium spp. and S. brevicaulis and 5) A. versicolor, C. arbuscula and S. brumptii. Comparison of Tables 1 and 3 show that some building materials are more prone to fungal growth after water damage and that some materials have a high fungal load (Table 3: high total column value), while others only support little fungal growth (Table 3: low total column value). For example, wallpaper often becomes fungal ridden (Table 1: 12.6 %) and with a high fungal load, where as glass fiber showed fungal growth less often (Table 1: 3.6 %) and had a low fungal load. Comparisons of Tables 3 and 4 show that some fungi have a high occurrence, but no association with particular materials (e.g. Penicillium spp.), while other fungi have a moderate occurrence and a specific association to particular materials (e.g. Stachybotrys spp./gypsum and Ulocladium spp./wallpaper). Others again have a low occurrence, but a tight association with a particular material (e.g. A. pullulans/wood [blue stain fungus] and Phoma spp./glass fiber). 293 294 295 296 297 298 299 300 301 302 303 Discussion Sampling methods There exists no single sampling method that can detect all fungal species in a moldy building, because all methods (e.g. air or surface sampling with either spore counting or cultivation) are, in one way or another, biased. Air sampling is unreliable because it favors fungi that produce large quantities of small, dry spores, such as Aspergillus spp., Cladosporium spp., and Penicillium spp. (38) and discriminates against fungi that produce small amounts of spores, large spores or spores in slime, such as Acremonium spp., Chaetomium spp., Stachybotrys spp., Trichoderma spp., and Ulocladium spp. Besides, fungal diversity is different in outdoor air (22), in indoor air (12) and in house dust (13, 52) compared with each other and with moldy building materials (14). Air sampling 13

304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 alone may give an incorrect picture of the fungal diversity actually present in a moldy building. In this study surface sampling with contact plates containing V8 medium with antibiotics was used. Other agar media (e.g. water agar, MEA or DG18 (35, 36) have been recommended in older literature, but new research (38) has shown that important toxigenic fungi like Chaetomium spp., Stachybotrys spp. and Trichoderma spp. do not grow or sporulate well on these media. V8, on the other hand, has been shown to support good growth and sporulation of most indoor fungi (3, 12, 22, 38) and V8 allows direct genus identification of indoor fungi and species identification of most Alternaria, Aspergillus, Cladosporium and Zygomycetes. A disadvantage with V8 is that it does not allow growth of dust fungi such as Eurotium spp. or Wallemia spp. (38) or identification to species level of Penicillium. A disadvantage with all sampling methods based on cultivation is that they only detect viable fungal spores. In the case of Stachybotrys spp. detection is essential, because dead spores coated in toxins might still be present on moldy materials. This can be overcome by using Sellotape (Scotch tape) to sample directly on the surface of moldy building materials. This method complements contact plates or swabs and detects the non-viable and non-culturable fungi (35, 38). The tape preparation method is cheap and quick and can be used on the sampling site prior to the V8 contact plates. Surface sampling using DNA detection would be ideal, but accurate DNA detection and identification of many filamentous fungi is not yet possible (38). None of these sampling methods would detect hyphal fragments, MVOCs or metabolites that might affect occupants living or working in these environments. 323 324 325 326 327 Fungal diversity Most studies on indoor fungi either deal with surveys of a few fungal species sampled on surfaces (e.g. 4) or many fungal genera detected using air or dilution sampling (e.g. 5 and 19, respectively). This study deals with many fungal genera (Table 2) on many surface components (Table 1). Our 14

328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 findings (Table 2) do not corroborate the findings reported in WHO guidelines (51). In this study Chaetomium spp., Acremonium spp., Sporothrix spp., Calcarisporium spp., and Scopulariopsis spp. were detected together with Arthrinium spp., Aureobasidium spp. and Gliocladium spp. These genera were not listed in the WHO guidelines (51). WHO reports that Eurotium spp. and Wallemia sebi are common, but these dust fungi were not detected in this study because V8 was used. Epicoccum spp. and Phialophora spp. were found in this study, but so rarely (< 0.5 %) that they were deleted prior to analyses. Our results, however, correspond well to the findings of Gravesen et al. (14), except for Stachybotrys spp. and Ulocladium spp., where our study found 3.9 % Stachybotrys spp. and 8.0 % Ulocladium spp. compared to 19 and 21 %, respectively, in Gravesen et al. (14). One reason could be that Gravesen et al. (14) used tape preparations on the moldy materials as a supplement to the V8 medium, which ensures that non-viable Stachybotrys spp. and Ulocladium spp. spores were also detected. Another reason could be that Gravesen et al. (14) examined more gypsum samples (11.1 %) than we did (7.7 %), with which Stachybotrys spp. and to some extent Ulocladium spp. are associated. Identification to species level is always recommended, in order to know the fungal diversity (mycobiota) before building renovation commences. Large amounts of alien spores from outside sources can be introduced to a building under renovation, and knowledge of the original mycobiota can be used as a control measure of renovation quality. Identification to species level is also important from a health perspective, since several fungal genera contain species capable of producing species specific metabolites, mycotoxins (9, 25, 28, 47, 53) and allergens (41, 50). Some fungal genera, however, are notoriously difficult to identify to species level without further cultivation, especially Penicillium. Identification of a set of Penicillia randomly sampled from other indoor samples showed that P. chrysogenum was the most common Penicillium, constituting more than 70 % of the cultivated Penicillia. Using this approximation P. chrysogenum would be the most 15

352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 common fungal species in moldy buildings with ca. 50 %, being detected twice as often as A. versicolor. Using the same estimation P. brevicompactum, P. citreonigrum, P. corylophilum, P. crustosum, P. olsonii, P. palitans and P. solitum would constitute 4-5 % each. Several other Penicillium spp. have been associated with indoor environments, such as P. citrinum, P. digitatum, P. expansum, P. glabrum, P. italicum and P. roqueforti, but only when air sampling has been used (e.g. 8, 40). Most of these Penicillium species are associated with foods, plants, and herbs (8, 38) and may be more associated with the presence of moldy food in the building than the moldy building materials themselves. Other fungal genera can be difficult to identify to species directly on V8. Newer literature, however, suggest that the most common Chaetomium and Acremonium species found on water damaged building materials are C. globosum and A. strictum, respectively (38). Ulocladium spp. are also common and studies have shown that several species, U. alternariae, U. atrum and U. oudemansii, can be detected on water damaged building materials (1). Alternaria tenuissima, on the other hand, might often be confused with U. chartarum, because both species produce spores in unbranched chains. The most common Trichoderma species in water damaged buildings belongs to the species complex of T. harzianum (38), whereas the only Stachybotrys species found are Stachybotrys chartarum and S. chlorohalonata (2). The findings presented in this study, that P. chrysogenum, A. versicolor and C. globosum are the three most common fungal species on water damaged building materials, correspond very well with the findings of Polizzi et al. (32). They repeatedly detected the mycotoxins, roquefortine C, sterigmatocystin and chaetoglobosin A, in air, dust, fungal biomass and wallpaper samples. Roquefortine C, sterigmatocystin and chaetoglobosin A are the major metabolites produced by P. chrysogenum, A. versicolor and C. globosum, respectively (38). Polizzi et al. (32) also detected roridin E in air samples and ochratoxin A and aflatoxins in air, dust and fungal biomass samples, 16

376 377 378 but were not able to detect the producing fungal species. These toxins are produced by Stachybotrys spp., A. niger, A. ochraceus, and A. flavus (38), which are also quite common in moldy buildings (Table 2). 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 Associations between fungi and building materials This study, which is the largest of its kind, shows that there exists an associated mycobiota on different building materials as there exists an associated mycobiota on different food types (38). The association between Acremonium spp., P. chrysogenum, Stachybotrys spp., Ulocladium spp. on gypsum and wallpaper was indicated by both ordination methods (PCA and CA) and the PCO/minimum spanning tree (Figs. 1 and 2 and Table 3). Production of neutral cellulases have been found in these fungi (31, 44) and may be a common ability of many indoor fungi that have found their niche on damp gypsum or plaster walls clad with wallpaper. A second strong association was seen between A. phaeospermum, A. pullulans, C. herbarum, Trichoderma spp. and yeasts on different types of wood and plywood in all three analyses (PCA, CA and PCO). Alternaria tenuissima, Fusarium spp., Gliocladium spp., R. mucilaginosa and T. pullulans were associated with the group in two out of three analyses. These fungi are also known for their production of neutral cellulases (44), but differ from fungi on damp walls in that they need higher water activity (39). The third association was seen between A. fumigatus, A. melleus, A. niger, A. ochraceus, Chaetomium spp., M. racemosus and M. spinosus on concrete and other floor related materials, such as linoleum, cork and the glue used to secure them. Concrete is mostly used to cast foundations, floors and other horizontal structures, and will hold soil, dirt, and dust when compared with vertical surfaces. These Aspergillus, Chaetomium and Mucor species are common in dust (13) and in hypersaline water and soil (16, 45). They may be introduced along with dirt and may tolerate alkaline conditions in the concrete, beginning to grow when the concrete gets wet. 17

400 401 402 403 404 405 406 407 408 409 410 The results presented here can aid the building profession in choosing materials, which are less susceptible to fungal growth (e.g. glass fiber instead of wallpaper) and by manufacturers of building materials to develop non-toxic, fungal or water resistant materials (e.g. coating of gypsum board against Stachybotrys spp.). Health authorities can use the results to facilitate the development of standardized allergen extracts to test for Type I allergy to specific indoor fungi: Acremonium strictum, Aspergillus versicolor, Chaetomium globosum, Stachybotrys chartarum, S. chlorohalonata and Ulocladium alternariae. The results can also be used by the scientific community to focus research on the physiology and ecology of toxigenic species, their potential production of mycotoxins, MVOCs and micro-particles on building materials, and to aid in the development of new and better detection and identification methods for fungal growth on building materials. 411 412 413 414 Acknowledgment The authors would like to thank Dr. Ulf Thrane, DTU for fruitful discussions and VILLUM FONDEN for financial support. 18

415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 References 1. Andersen, B. and M. Hollensted. 2008. Metabolite production by different Ulocladium species. Int. J. Food. Microbiol., 126:172-179. 2. Andersen, B., K.F. Nielsen, U. Thrane, T. Szaro,, J.W. Taylor, and B.B. Jarvis. 2003. Molecular and phenotypic descriptions of Stachybotrys chlorohalonata sp. nov. and two chemotypes of Stachybotrys chartarum found in water-damaged buildings. Mycologia 95:1227-1238. 3. Andersen, B., and A.T. Nissen AT. 2000. Evaluation of media for detection of Stachybotrys and Chaetomium species associated with water-damaged buildings. Int. Biodeter. Biodegr. 46:111-116. 4. Bellanger, A.P., G. Reboux, S. Roussel, F. Grenouillet, E. Didier-Scherer,J.C. Dalphin, and L. Millon. 2009. Indoor fungal contamination of moisture-damaged and allergic patients housing analysed using real-time PCR. Lett. Appl. Microbiol. 49:260-266. 5. Chao, H.J., J. Schwartz, D.K. Milton, and H.A. Burge. 2002. Populations and determinants of airborne fungi in large office buildings. Environ. Health Perspect. 110:777-782. 6. de Hoog, G.S., J. Guarro, J. Gené, and M.J. Figueras. 2000. Atlas of Clinical Fungi. Centraalbureau voor Schimmelcultures - KNAW. Utrecht, The Netherlands. 7. Domsch, K.H., W. Gams, and T.-H. Anderson. 1980. Compendium of soil Fungi. First edition. IHW-Verlag. Eching bei München, Germany. 8. Frisvad, J.C., and S. Gravesen. 1994. Mycotoxins of indoor Penicillium and Aspergillus. p. 281-290. In Samson RA, Flannigan B, Flannigan ME, Verhoeff AP, Adan OCG, (eds.), Health Implications of Fungi in Indoor Environments. Air Quality Monographs Vol. 2. Elsevier, Amsterdam.. 19

439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 9. Frisvad, J.C., B. Andersen, and U. Thrane. 2008. The use of secondary metabolite profiling in chemotaxonomy of filamentous fungi. Mycol. Res. 112:231-240. 10. Garrett, M.H., P.R. Rayment, M.A. Hooper, M.J. Abramson, and B.M. Hooper. 1998. Indoor airborne fungal spores, house dampness and associations with environmental factors and respiratory health in children. Clin. Exp. Allergy. 28:459-467. 11. Grant, C., C.A. Hunter, B. Flannigan, and A.F. Bravery. 1989. The moisture requirements of moulds isolated from domestic dwellings. Int. Biodeter. 25:259-284. 12. Gravesen S. 1972. Identification and quantification of indoor airborne micro-fungi during 12 months from 44 Danish homes. Acta Allergol. 27:337-354. 13. Gravesen S. 1978. Identification and prevalence of culturable mesophilic microfungi in house dust from 100 Danish homes. Allergy 33:268-272. 14. Gravesen, S., P.A. Nielsen, R. Iversen, and K.F. Nielsen. 1999. Microfungal contamination of damp buildings examples of risk constructions and risk materials. Environ. Health Perspect. 107:505-508. 15. Greenacre, M.J. 1984. Theory and applications of correspondence analysis. Academic Press, London. 16. Grishkan, I., Nevo, E., and S.P. Wasser. 2003. Soil micromycete diversity in the hypersaline Dead Sea costal area, Israel. Mycol. Prog. 2:19-28. 17. Holme, J., L. Hägerhed-Engman, J. Mattsson, J. Sundell J, and C.-G. Bornehag. 2010. Culturable mold in indoor air and its association with moisture-related problems and asthma and allergy among Swedish children. Indoor Air 20:329-340. 18. Horner, W.E., A. Helbling, J.E. Salvaggio, and S.B. Lehrer. 1995. Fungal allergens. Clin. Microbiol. Rev. 8:161-179. 19. Hyvärinen, A., T. Meklin, A. Vepäläinen, and A. Nevalainen. 2002. Fungi and 20

463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 actionobacteria in moisture-damaged building materials - concentrations and diversity. Int. Biodeter. Biodegr. 49:27-37. 20. Jarvis, J,Q, and R.R. Morey. 2001. Allergic respiratory disease and fungal remediation in a building in a subtropical climate. Appl. Occup. Environ. Hyg. 16:380-388. 21. Kildesø, J., H. Würtz, K.F. Nielsen, K.F., P. Kruse, K. Wilkins, U. Thrane, S. Gravesen, P.A. Nielsen, and T. Schneider. 2003. Determination of fungal spore release from wet building materials. Indoor Air 13:148-155. 22. Larsen, L., and S. Gravesen. 1991. Seasonal variation of outdoor airborne viable microfungi in Copenhagen, Denmark. Grana 30:467-471. 23. Lee, T.G. 2003. Health symptoms caused by molds in a courthouse. Arch. Environ. Health. 58:442-446. 24. Miller, J.D., Sun, M., Gilyan, A., Roy, J., and T.G. Rand. 2010. Inflammation-associated gene transcription and expression in mouse lungs induced by low molecular weight compounds from fungi from the built environment. Chem.-Biol. Interact.183:113-124. 25. Mitchell, C.S., J. Zhang, T. Sigsgaard, M. Jantunen, P.J. Lioy, R. Samson, and M.H. Karol. 2007. Current state of the science: health effects and indoor environmental quality. Environ. Health Perspect. 115:958-964. 26. Nielsen, K.F. 2002. Mould growth on building materials: Secondary metabolites, mycotoxins and biomarkers. PhD thesis.technical University of Denmark, Lyngby. 27. Nielsen, K.F., G. Holm, L.P. Uttrup, and P.A. Nielsen. 2004. Mould growth on building materials under low water activities. Influence of humidity and temperature on fungal growth and secondary metabolism. Int. Biodeter. Biodegr. 54:325-336. 28. Nielsen, K.F., U. Thrane, T.O. Larsen, P.A. Nielsen, and S. Gravesen. 1998. Production of mycotoxins on artificially inoculated building materials. Int. Biodeter. Biodegr. 42:9-16. 21

487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 29. Niemeier, R.T., S.K. Sivasubramani, T. Reponen T, and S.A. Grinshpun. 2006. Assessment of fungal contaminatin in moldy homes: Comparison of different methods. J. Occup. Environ. Hyg. 3:262-273. 30. Park, D. 1982. Phylloplane fungi: Tolerance of hyphal tips to drying. Trans. Brit. Mycol. Soc. 79:174-178. 31. Picart, P., P. Diaz, and F.I.J. Pastor. 2008. Stachybotrys atra BP-A produces alkaliresistant and thermostable cellulases. Antonie van Leeuwenhoek 94:307-316. 32. Polizzi, V., B. Delmulle, A. Adams, A. Moretti, A. Susca, A.M. Picco, Y. Rosseel, R. Kindt, J. Van Bocxlaer, N. De Kimpe C. Van Peteghem, and S. de Saeger. 2009. JEM spotlight: Fungi mycotoxins and microbial volatile organic compounds in mouldy interiors from water-damaged buildings. J. Environ. Monitor. 11:1849-1858. 33. Rand, T.G., S. Giles, J. Flemming, J.D. Miller, and E. Puniani. 2005. Inflammatory and cytotoxic responses in mouse lungs exposed to purified toxins from building isolated Penicillium brevicompactum Dierckx and P. chrysogenum Thom. Toxicol. Sci. 87:213-222. 34. Rand, T.G., M. Sun, A. Gilyan, J. Downey, and Miller JD. 2010. Dectin-1 and inflammation-associated gene transcription and expression in mouse lungs by a toxic (1,3)- β-d-glucan. Arch. Toxicol. 84:205-220. 35. Reenen-Hoekstra, S.E. van, R.A. Samson, A.P. Verhoeff, J.H. van Wijnen, and B. Brunekreef. 1991. Detection and identification of moulds in Dutch houses and nonindustrial working environments. Grana 30:418-423. 36. Samson, R.A., and E.S. Hoekstra. 1994. Common fungi occurring in indoor environments. p. 541-546. In Samson, R.A., B. Flannigan, M.E. Flannigan, A.P. Verhoeff, and O.G.C. Adan (Eds.), Health Implications of Fungi in Indoor Environments. Air Quality Monographs Vol. 2, Elsevier, Amsterdam. 22

511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 37. Samson, R.A., E.S. Hoekstra, J.C. Frisvad, and O. Filtenborg O. (eds.). 2002. Introduction to food- and airborne fungi. Sixth edition. Centraalbureau voor Schimmelcultures, CBS KNAW. Utrecht, The Netherlands. 38. Samson, R.A., J. Houbraken, U. Thrane, J.C. Frisvad JC, Andersen B. 2010. Food and Indoor Fungi. First Edition. Centraalbureau voor Schimmelcultures, CBS KNAW. Utrecht, The Netherlands. 39. Sautour, M., C. Soares Mansur, C. Divies C, M. Bensoussan, M., and P. Dantigny. 2002. Comparison of the effects of temperature and water activity on growth rate of food spoilage moulds. J. Ind. Microbiol. Biot. 28:311-315. 40. Sen, B, and A. Asan. 2009. Fungal flora in indoor and outdoor air of different residential houses in Tekirdag City (Turkey): seasonal distribution and relationship with climatic factors. Environ. Monit. Assess. 151:209-219. 41. Simon-Nobbe, B., U. Denk, V. Pöll, R. Rid R, and M. Breitenbach. 2008. The spectrum of fungal allergy. Int. Arch. Allergy Immunol. 145:58-86. 42. Sivasubramani, S.K., R.T. Niemeier, T. Reponen, and S.A. Grinshpun. Assessment of the aerosolization potential for fungal spores in mouldy homes. Indoor Air 14:405-412. 43. Sneath, P.H.A., and R.R. Sokal. 1973. Numerical Taxonomy, W.H. Freeman and Co., San Francisco. 44. Solovyeva, I.V., O.N. Okunev, E.G. Kryukova, and G.A. Kochkina. 1999. Occurrence of neutral and alkaline cellulases among alkali-tolerant micromycetes. Syst. Appl. Microbiol. 22:546-550. 45. Steiman, R., L. Ford, V. Ducros, J.-L. Lafond, P. Guiraud. 2004. First survey of fungi in hypersaline soil and water of Mono Lake area (California). Antonie van Leeuwenhoek 85:69-83. 23

535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 46. Tucker, K., J.L. Stolze, A.H. Kennedy, and N.P. Money. 2007. Biomechanics of conidial dispersal in the toxic mold Stachybotrys chartarum. Fungal Genet. Biol. 44:641-647. 47. Tuomi, T., K. Reijula T. Johnsson, E.L. Hemminki, E.L. Hintikka, O. Lindroos, S. Kalso, P. Koukila-Kähkölä, H. Mussalo-Rauhamaa H, and T. Haahtela. 2000. Mycotoxins in crude building materials from water-damaged buildings. Appl. Environ. Microbiol. 66:1899-1904. 48. Verhoeff, A.P., J.H. van Wijnen, E.S. van Reenen-Hoekstra, R.A. Samson, R.T. van Strien, and B. Brunekreef. 1994. Fungal propagules in house dust II. Relation with residential characteristics and respiratoral symptoms. Allergy 49:540-547. 49. Viitanen, H., J. Vinha, K. Salminen, T. Ojanen, R. Peuhkuri, L. Paajanen, K. Lähdesmäki. 2010. Moisture and bio-deterioration risk of building materials and structures. J. Build. Physics 33:201-224. 50. Vijay, H.M., M. Abebe, V. Kumar, M. DeVouge, T. Schrader, A. Thaker, P. Comtois, and B. Escamilla-Garcia. 2005. Allergenic and mutagenic characterization of 14 Penicillium species. Aerobiologia 21:95-103. 51. WHO Regional Office for Europe. 2009. WHO guidelines for indoor air quality: dampness and mould. (www.euro.who.int/ data/assets/pdf_file/0017/43325/e92645.pdf, accessed 16 July 2010). 52. Wickman, M., S. Gravesen, S.L. Nordvall, G. Pershagen G, and J. Sundell.1992. Indoor viable dust-bound microfungi in relation to residential characteristics, living habits, and symptoms in atropic and control children. J. Allergy Clin. Immunol. 89:752-759. 53. Wålinder, R., L. Ernstgård, G. Johanson, D. Norbäck, P. Venge, and Wieslander, G. 2005. Acute effects of a fungal volatile compound. Environ Health. Perspect. 113:1775-1778. 24

559 560 561 562 563 564 565 566 Figure legends Fig. 1. Loadings plot from the Principal Component Analysis (PCA) based on the qualitative Matrix A [5,353 samples (30 materials and 42 fungi)]. The plot shows associations between building materials and fungi (e.g. between Wood and A. tenuissima, C. herbarum, R. mucilaginosa and yeasts). Fungi and building materials encircled are particularly associated. Fungi or components close to the centroid have little or no association with each other, occur infrequently (< 4%) and have little or no influence on the PCA model. Axes are principal components, PC 1 and PC 2, with loading values. 567 568 569 570 571 572 573 574 Fig. 2. Biplot from the Correspondence Analysis (CA) based on the quantitative sum Table B (Matrix B [4,241 samples (25 materials and 41 fungi)] summarized in a 41 25 table). The biplot shows the association between building material and fungal identity (e.g. between gypsum, glass fiber, wallpaper, Ulocladium spp. and Stachybotrys spp.). The dotted lines show the distance of any fungus to the centroid, i.e. fungi farthest away from the centroid deviate the most from what would be expected based on the whole data table. Axes are correspondence components, DIM 1 and DIM 2, with score and loading values. 25

1 2 3 Table 1. Qualitative (qual) and quantitative (quan) frequencies of building material components with fungal growth from water damaged buildings. An additional 18 building materials occurred in less than 2 % each. A sample may contain more than one material component. Frequency (%) Material component qual (n=5,353) quan (n=4,241) Plaster 23.5 22.1 Concrete 19.0 24.1 Wood 18.4 19.0 Wallpaper 12.6 6.3 Gypsum 7.7 8.1 Paint 5.2 5.8 Mineral wool 3.8 3.3 Glass fiber 3.6 3.6 Plywood 3.0 4.5 Brick 2.1 2.5 Chipboard 2.0 2.5 Linoleum 2.0 3.5 1

4 5 6 7 8 Table 2. Qualitative (qual) and quantitative (quan) frequencies of fungal species and genera on water damaged building materials. An additional 15 fungi occurred in less than 2 % of the samples. A sample may contain more than one fungus. Frequency (%) Fungi qual (n=5,353) quan (n=287,169) Penicillium spp. 69.5 39.7 Aspergillus versicolor 26.5 15.6 Chaetomium spp. 16.5 3.1 Acremonium spp. 14.9 7.8 Ulocladium spp. 8.0 2.1 Cladosporium sphaerospermum 7.4 4.9 Mucor plumbeus (syn. M. spinosus) 7.2 0.3 Trichoderma spp. 6.7 0.4 C. herbarum 6.6 1.5 Alternaria tenuissima 6.5 0.7 Sporothrix spp. 6.4 3.3 A. niger 6.1 0.5 Yeasts 5.1 2.6 Rhodotorula mucilaginosa 5.1 2.3 A. ochraceus 5.0 0.9 P. chrysogenum * (syn. P. notatum) 4.5 2.5 Rhizopus stolonifer (syn. R. nigricans) 4.1 0.1 Stachybotrys spp. 3.9 1.9 A. fumigatus 3.8 0.2 Aspergillus spp. 3.6 1.2 Mucor spp. 3.3 0.2 Mycelia sterilia 3.1 - A. wentii 3.1 1.3 Calcarisporium arbuscula 2.7 1.2 Scopulariopsis brevicaulis 2.1 0.5 Fusarium spp. 2.0 0.4 M. racemosus 2.0 0.1 *Underestimated, as several of the Penicillium spp. may also be P. chrysogenum. 2

9 10 11 Table 3. Qualitative (qual) and quantitative (quan) distribution (%) of associated fungi on different building materials. Values in bold give the materials on which the fungus is overrepresented (the associated mycobiota). Row sums are 100%. Concrete Glass fiber Gypsum Plaster Plywood Wallpaper Wood Fungi qual quan qual quan qual quan qual quan qual quan qual quan qual quan Acremonium spp. 13 14 3 3 8 8 25 22 3 5 8 6 14 15 Aspergillus niger 32 35 2 0 7 5 11 10 1 0 4 1 18 23 A. ochraceus 33 52 2 4 4 3 20 6 1 0 5 5 12 8 A. versicolor 20 26 3 3 4 4 25 21 1 1 11 7 11 10 Aureobasidium pullulans 3 0 0 9 6 0 14 11 9 6 3 0 46 37 Chaetomium spp. 25 35 2 2 6 9 15 11 1 2 10 3 17 18 Cladosporium herbarum 11 15 1 1 5 5 14 15 5 9 6 3 24 27 C. sphaerospermum 8 10 4 3 3 4 22 24 6 11 12 6 17 16 M. racemosus 36 54 1 0 3 3 17 14 0 0 5 1 15 10 Paecilomyces variotii 13 8 0 1 13 15 10 8 19 15 2 3 27 17 Penicillium spp. 19 26 3 2 6 6 20 16 2 3 10 6 17 18 P. chrysogenum 12 18 2 2 11 14 21 18 3 4 16 5 13 15 Phoma spp. 9 12 9 12 4 1 15 9 2 4 12 6 12 11 Sporothrix spp. 18 21 3 2 3 4 39 37 0 2 7 4 7 7 Stachybotrys spp. 10 4 2 10 25 39 22 13 1 0 10 8 9 7 Trichoderma spp. 14 16 2 2 9 9 15 13 4 8 6 2 25 30 Ulocladium spp. 9 3 5 2 7 15 21 34 1 1 29 16 12 9 Yeasts 8 7 2 3 8 5 10 11 7 9 5 2 28 39 3

12 13 14 Table 4. Qualitative (qual) and quantitative (quan) distribution (%) of dominant fungi on particular building materials. Values in bold give the three most dominant fungi. Column sums are 100%. Concrete Glass fiber Gypsum Plaster Plywood Wallpaper Wood Fungi qual quan qual quan qual quan qual quan qual quan qual quan qual quan Acremonium spp. 4 5 8 9 7 9 8 10 8 10 5 8 5 7 Aspergillus niger 4 1 2 0 3 0 1 0 1 0 1 0 3 1 A. ochraceus 3 2 2 1 1 0 2 0 1 0 1 1 1 0 A. versicolor 12 19 12 19 6 10 14 19 5 6 13 20 7 10 Aureobasidium pullulans 0 0 0 1 0 0 0 0 1 0 0 0 1 0 Chaetomium spp. 9 5 5 2 6 4 5 2 3 1 7 1 7 3 Cladosporium herbarum 2 1 1 1 2 1 2 1 5 3 2 1 4 2 C. sphaerospermum 1 2 4 5 2 3 3 7 7 13 4 5 3 5 M. racemosus 1 0 0 0 0 0 1 0 0 0 0 0 1 0 Paecilomyces variotii 0 0 0 0 1 1 0 0 2 2 0 0 1 1 Penicillium spp. 28 46 26 27 27 34 27 35 31 41 30 39 28 44 P. chrysogenum 1 2 2 2 3 5 2 2 1 2 3 2 1 2 Phoma spp. 0 0 3 3 1 0 1 0 1 1 1 1 1 1 Sporothrix spp. 3 3 3 2 1 2 5 7 0 1 2 3 1 1 Stachybotrys spp. 1 0 1 7 6 12 2 1 1 0 2 3 1 1 Trichoderma spp. 2 0 2 0 4 1 2 0 4 1 2 0 4 1 Ulocladium spp. 2 0 6 2 4 5 3 4 2 1 10 6 2 1 Yeasts 1 1 1 2 3 2 1 2 6 6 1 1 4 6 4

0.3 PC2 X-loadings Fig. 1 0.2 0.1 0-0.1 Wood Rho mucilaginosa Yeast Alt tenuissima Cla herbarum Phoma spp Cla sphaerospermum Ulocladium spp Wallpaper Plaster Acremonium spp Paint Plywood Grout Glass fiber Pen chrysogenum Gypsum Aur pullulans Stachybotrys sp Fusarium spp Gliocladium spp Mycelia sterili Trn pullulans Felt Gas concrete Trichoderma Cardboard Brick spp Art phaeospermumbitumen AlternariaTextile Eternit spp MDF Filler fibre board Pae variotii Mucor spp Pae lilacinus Mineral Paper Masonite wool Wood-wool cemen Cork Tar-treated Ver albo-atrum Vapour barrier Aspergillus spp Sco brumptii Carpet Vinyl Absidia spp Sporothrix spp Rhi nigricans Chipboard Asp sydowii Sco brevicaulis Asp ustus Glue Asp wentii Muc racemosus Asp versicolor Linoleum Cal arbuscula Asp flavus Asp melleus Asp fumigatus -0.2-0.3 Muc spinosus Penicillium spp Chaetomium spp Asp niger Asp ochraceus -0.4 Concrete -0.5-0.3-0.2-0.1 0 0.1 0.2 0.3 5353-42-30 PC1

0.52 Ulocladium spp Stachybotrys spp Fig 2 Cal arbuscula Vinyl Cork Sco brumptii Sporothrix spp Asp sydowii Wallpaper/Glass fiber Mucor spp Asp flavus Ver albo-atrum Asp ochraceus MDF fiber Asp versicolor Gypsum Wallpaper Muc racemosus Plaster 0.16 Aspergillus spp Glue Gas concrete Brick Glass fiber Asp fumigatus Concrete Asp wentii Trn pullulans Chip board Pen chrysogenum Paper Asp melleus Acremonium spp Chaetomium spp Muc spinosus Alt tenuissima Linoleum Penicillium spp Asp niger Asp ustus Paint Absidia spp Masonite Pae lilacinus -0.20 Cardboard Alternaria spp Bitumen Fusarium spp Troltex Phoma spp Cla sphaerospermum Insulation Sco brevicaulis DIM-2 Rhi nigricans -0.56 Wood Vapour barrier Trichoderma spp Grout/tile Cla herbarum Pae variotii Aur pullulans Art phaeospermum Rho mucilaginosa Plywood Yeast 4241-41-25-0.92 Gliocladium spp DIM-1-0.72-0.29 0.15 0.58 1.01